Numeral System and Its

Importance
HONG@IS.NAIST.JP
ROOM: B405
 Contents
 Number System Introduction
Number systems used by human
Number systems used by computer

Number System Conversion

Signed Number Representation
 Fractional Number Representation
Fixed point number
Floating point number
 Size of Number: Complexity vs. Performance
2
 Number systems used by human
 Number systems is a writing system for expressing number
 Number system is necessary for solving mathematics problem

Simple natural number: 1,2,3,…,10

https://www.youtube.com/watch?v=cy-8lPVKLIo
3
 Number systems used by human

 For business activities  Whole number and Natural number are developed, ex: 0,1,2,305, etc.

4
 Number systems used by human

 To share the foods  Fractional number have been developed, ex: 1/5, 1/2, etc.

5
 Number systems used by human
𝜋𝑅2

𝑅 1 √𝟐

 For building house, pyramid, complex

architecture, etc.
 Science discovery activities: discover
the universe, stars, moon, etc.
 Irrational number, real number,
complex number have developed.

6
 Number systems used by human
a,b: real numbers a+jb Complex Number 1 + j2, 1.54 – j0.3, etc
j: imaginary unit (j2 = -1)

Fractional number Real Number Imaginary Number

1/2, 0.35, 𝜋 = 3.141592…

5, 3/5, etc
Rational Number Irrational Number 2 = 1.41421 …
integer
Integer

Whole number

Natural Number

7
Number systems used by Computer
 Electronic circuit or computer operates based on the on-off switch of
electric current.  The basic number system of computer is BINARY
number. ON  1
OFF  0

 One binary number is called as a BIT

 A long chain of BITs may be difficult to remembered for computer
engineers  OCTAL and HEXA-DECIMAL number systems are more
convenient.
Binary 100111010110 Binary 100111010110
Octal 4 7 2 6 Hexa-decimal 9 B 6
(100111010110)2 = (4726)8 (100111010110)2 = (9B6)16
8
Number systems used by Computer
 For computer users, who do not familiar with computer architecture
 Binary, Octal, and Hexa-decimal are difficult for them
 Decimal System (base-10) is used.
 SUMMARY:

System Base Digits

Binary 2 0, 1
Octal 8 0, 1, 2, 3, 4, 5, 6, 7
Hexa-decimal 16 0, 1, 2, …, 9, A, B, C, D, E, F
Decimal 10 0, 1, 2, …, 9
9
Contents
 Number System Introduction
Number systems used by human
Number systems used by computer

 Number System Conversion

 Signed Number Representation
 Fractional Number Representation
Fixed point number
Floating point number
 Size of Number: Complexity vs. Performance
10
 Number System Representation
 A number has two parts: Integer & Fractional parts
 Position notation of number N = (an-1 ··· a0 . a-1 ··· a-m)r
1 0 -1 -2
Example: (13.25)10  (r = 10, a1 = 1, a0 = 3, a-1 = 2, a-2 = 5)
(1001.1101)2 , etc. n 1

 Polynomial notation of number:

N  i
a r
i  m
i
 a n 1  r n 1
   a m  r m

i = 1 0 -1 -2
Example: N = (13.25)10 can be written as:
 N = 1×101 + 3×100 + 2×10-1 + 5×10-2
i = 2 1 0 -1-2-3-4
N = (101.1101)2 can be rewritten as:
 N = 1×22 + 0×21 + 1×20 + 1×2-1 + 1×2-2 + 0×2-3 +1×2-4
 N = 1×22 + 1×20 + 1×2-1 + 1×2-2 + 1×2-4

11
 Exercise 4-1
 Write the polynomial notation of the following numbers:
N = (101.1101)2  N = ?
N = (765.02)8  N = ?
N = (25.5)10  N = ?

12
 Exercise 4-1: Answer
i = 2 1 0 -1-2-3-4
N = (101.1101)2  N = 1×22 + 1×20 + 1×2-1 + 1×2-2 + 1×2-4
Or, N = 1×22 + 0×21 + 1×20 + 1×2-1 + 1×2-2 + 0×2-3 + 1×2-4

i = 2 1 0 -1-2
N = (765.02)8  N = 7×82 + 6×81 + 5×80 + 0×8-1 + 2×8-2

i = 1 0 -1
N = (25.5)10  N = 2×101 + 5×100 + 5×10-1

13
Number conversions

Convert from base-B to base-10 (B can be 2, 8, 16):

Convert from base-10 to base-B
 Convert the Integer part: Radix Divide method
 Convert the Fractional part: Radix Multiplier method

14
 Number conversions
Convert from base-B to base-10 (B can be 2, 8, 16):
Two steps: n 1
◦ Step 1: Write the polynomial notation of base-B number N 
i  m
ai r i 
◦ Step 2: Calculate the polynomial using base-10 arithmetic

Example-1: (101.1)2  (??)10

Step 1: (101.1)2 = 1×22 + 0×21 + 1×20 + 1×2-1
Step 2: (101.1)2 = (4)10 +(0)10 + (1)10 + (0.5)10 = (5.5)10  (101.1)2 = (5.5)10

Example-2: (9e)16  (??)10

Step 1: (9e)16 = 9×161 + 14×160
Step 2: (9e)16 = (144)10 + (14)10 = (158)10  (9e)16 = (158)10

15
 Exercise 4-2
 Convert the following numbers into base-10:
 (11011011)2 = (??)10
 (25)8 = (??)10

16
 Exercise 4-2: Answer
(11011011)2 = (??)10
 (11011011)2 = 1×27 + 1×26 + 1×24 + 1×23 + 1×21 + 1×20
= 128 + 64 + 16 +8 +2 +1
= (219)10
(25)8 = (??)10
 (25)8 = 2×81 + 5×80 = 16 + 5 = (21)10

17
 Number conversions
 Convert from base-10 to base-B
integer fraction
(N)10 = (A)B

1. Convert the integer part:

use Radix Divide Method
(13)10 = (?)2 (B=2)
Q(0) = 13 Q(2)/2 = 3/2 = 1, remainder:1
Q(0)/2 = 13/2 = 6, remainder: 1  a(3) = 1
 a(0) = 1  Q(3) = 1
 Q(1) = 6
Q(1)/2 = 6/2 = 3, remainder: 0 Q(3)/2 = 1/2 = 0, remainder: 1
 a(0) = 0  a(4) = 1
 Q(2) = 3  Q(4) = 0  completed!
 (13)10 = (1101)10 18
 Number conversions
 Convert from base-10 to base-B
integer fraction
(N)10 = (A)B

2. Convert the fraction part:

Use Radix Multiply method

(0.6875)10 = (?)2 (B=2)

F(0) = 0.6875 0.6875×2 = 1.375  a(0) = 1
F(1) = 0.375 0.375×2 = 0.75  a(1) = 0
F(2) = 0.75 0.75×2 = 1.5  a(2) = 1
F(3) = 0.5 0.5×2 = 1.0  a(3) = 1
F(4) = 0  completed  (0.6875)10 = (0.1011)2
19
 Number conversions
 Quick method to convert from binary to base-2x (octal x = 3, hexa x = 4) :
 Step-1: Divide binary number into sets of x digits, add leading zeros as needed.
Ex: 1011100  001 011 100 (octal case)
 0101 1100 (hexadecimal case)
 Step-2: Use the below tables to convert each x-digits to a single Binary Hexadecimal
octal/hex digit
Binary Octal Binary Hexadecimal 1000 8
0000 0 1001 9
000 0
Examples: 1010 A (or a)
001 1 0001 1
(001 011 100)2  (134)8
1011 B (or b)
(0101 1100)2  (5c)16 010 2 0010 2
011 3 0011 3 1100 C (or c)

0100 4 1101 D (or d)

100 4
0101 5 1110 E (or e)
101 5
0110 6 1111 F (or f)
110 6
111 7 0111 7 20
 Number conversions
 Quick method to convert from base-2x to binary (octal x = 3, hexa x = 4) :
 Step-1: Use below table to convert each octal/hexa digit into x-digit of binary.
 Step-2: Delete the left-most zero digits Binary Hexadecimal
Binary Hexadecimal 1000 8
Binary Octal 0000 0 1001 9
Examples: 0001 1 1010 A (or a)
000 0
 (35)8  (011 101)2
0010 2 1011 B (or b)
 (011 101)2  (11101)2 001 1
010 2 0011 3 1100 C (or c)
 (2F)16  (0010 1111)2 1101 D (or d)
011 3 0100 4
 (0010 1111)2 
100 4 0101 5 1110 E (or e)
(101111)2
101 5 0110 6 1111 F (or f)
110 6 0111 7
111 7
21
Exercise 4-3
 Convert the following numbers
 (58)16 = (?)2
 (1011010)2 = (?)16 = (?)8

22
 Exercise 4-3: Answer

◦ (58)16 = (0101 1000)2

◦ (0101 1010)2 = (5a)16
◦ (001 011 010)2 = (132)8

23
Contents
 Number System Introduction
Number systems used by human
Number systems used by computer

 Number System Conversion

 Signed Number Representation
 Fractional Number Representation
Fixed point number
Floating point number
 Size of Number: Complexity vs. Performance
24
 How to present signed number ?
 In mathematics, negative number is represented by minus “-
” sign.
 In computer hardware, numbers are represented as sequence
of bits (0, 1)  how to represent negative number?
 Signed Number Representation Methods:
Signed magnitude representation
Ones’ complement
Two’s complement

25
 Signed number Representation
- Signed magnitude representation-
Signed magnitude representation:
 Sign bit: commonly represented by the most significant bit
(MSB) (0 mean positive, 1 means negative).
 Magnitude: represented by the remained bits. It shows the
absolute value of the number. Sign
bit magnitude

MSB 0

Problem: Zero value can be represented as 0 0000 or 1 0000

Example: (0101)2 = (5)10 ; (1101)2 = (-5)10

26
 Signed number Representation
- Ones’ Complement -
 The ones’ complement form of a negative binary number is the bitwise NOT of its positive
number. Ex: (5)10 = (0101)2  (-5)10 = (1010)2
 If using N bits to represent a number  its range will be -(2N-1 - 1) to (2N-1 - 1)
 Steps to add two numbers:
 Do conventional binary addition
 Add an end-around carry
 Problems:
 Multiple zeros (0000 or 1111)
 End-around carry
Example:

27
 Signed number Representation
- Two’s Complement (1/2) -
 To avoid the end-around carry and multiple zeros problem, the two’s
complement has been introduced and widely used.
 The negative number is done by inverting all the bits and adding 1 to
the result.
 There is only one zero: 00000
 Method to find the negation of a number in two’s complement

 Method 1:
 Invert all the bits
 Add the result to 1
Example: Find (-3)10 = (?)2 using 4 bits
 Invert: 0011 (3)  1100
 Add 1: 1100 + 1 = 1101 (-3)
 Method 2:
 Find the first ‘1’ (start from right)
 Invert all the bits in the left of that ’1’.
Example: Find (-3)10 = (?)2
 Find the first ‘1’: 0011 (3)
 Invert all the left bits: 0011  1101 (-3)

28
Signed number Representation
- Two’s Complement (2/2) -
 Add two number in two’s complement format:
 Perform the binary adding
 Ignore the last carry bit.
 For example: Perform the following additions using 4 bits

1 0 1 0 (-6) 0 0 1 1 (3)
+ 0 1 0 1 (5) + 1 1 1 0 (-2)
----------- -----------
1 1 1 1 (-1) 1 0 0 0 1 (1)

29
 Exercise 4-4
 Find the signed magnitude representation of the following
numbers (using 5 bits):
 if (7)10 = (00111)2  (-7)10 = (?)2
 if (10)10 = (01010)2  (-10)10 = (?)2
 Find the ones’ complement of the following numbers (using 5 bits)
 (-7)10 = (?)2 (-10)10 = (?)2
 Find the negation of the following two’s complement numbers:
 0111010 1010011 110111010

30
 Exercise 4-4: Answer
 Find the signed magnitude representation of the following
numbers (using 5 bits):
 if (7)10 = (00111)2  (-7)10 = (10111)2
 if (10)10 = (01010)2  (-10)10 = (11010)2
 Find the ones’ complement of the following numbers (using 5 bits)
 (-7)10 = (11000)2 (-10)10 = (10101)2
 Find the negation of the following two’s complement numbers:
 0111010 1010011 110111010
 1000110 0101101 001000110

31
Contents
 Number System Introduction
Number systems used by human
Number systems used by computer

 Number System Conversion

 Signed Number Representation
 Fractional Number Representation
Fixed point number
Floating point number
 Size of Number: Complexity vs. Performance
32
 Fixed Point Number
Fixed-point number represents a real number having a fixed number of
digits after the radix point
UNSIGNED NUMBER SIGNED NUMBER

Radix point Radix point

m bits f bits sign m bits f bits
bit
Integer part Fractional part Integer part Fractional part

t=m+f t=m+f+1

m: number of integer bits

For examples: m = 3 bits, f = 4 bits, t = 7 bits f: number of fractional bits
 (0101100)2 = (010.1100)2 = (2.75)10 t: Total number of bits

33
 Floating Point Number
 Floating point number is a formula representation of a real number in trade-
off between Range and Precision
𝒔𝒊𝒈𝒏𝒊𝒇𝒊𝒄𝒂𝒏𝒅 × 𝒃𝒂𝒔𝒆−𝒆𝒙𝒑𝒐𝒏𝒆𝒏𝒕
For example: 𝟏𝟒𝟓𝟐 × 𝟏𝟎−𝟑 = 𝟏. 𝟒𝟓𝟐 𝟏𝟒𝟓𝟐 × 𝟏𝟎−𝟐 = 𝟏𝟒. 𝟓𝟐
 It means that the radix point can “float”.
 In computer, only base 2 can be used.

sign Exponent significand

MSB LSB
1 bit a bits b bits

t = a + b + 1 bits

For example: a = 3, b = 4, t = 8  (10100011)2 = -3×2-2 = (-0.75)10

2 3
34
Exercise 4-5
 Find the equivalent decimal values of the following
unsigned fixed point numbers (m = 4, f = 4, t = 8):
 (10011100)2 = (?)10
 Find the equivalent decimal values of the following
signed fixed point numbers (m = 3, f = 4, t = 8):
 (10011100)2 = (?)10
 Find the equivalent decimal values of the following
floating point numbers (a = 3, b = 4, t = 8):
 (00110101)2 = (?)10

35
Exercise 4-5: Answer
 Find the equivalent decimal values of the following
unsigned fixed point numbers (m = 4, f = 4, t = 8):
 (1001.1100)2 = (9.75)10
 Find the equivalent decimal values of the following
signed fixed point numbers (m = 3, f = 4, t = 8):
 (1 001.1100)2 = (- 1.75)10
 Find the equivalent decimal values of the following
floating point numbers (a = 3, b = 4, t = 8):
 (0 011 0101)2 = (+ 5×2-3)10 = (0.625)

36
Contents
 Number System Introduction
Number systems used by human
Number systems used by computer

 Number System Conversion

 Signed Number Representation
 Fractional Number Representation
Fixed point number
Floating point number
 Size of Number: Complexity vs. Performance
37
 Size of Number: Complexity vs. Performance
 Memory space: each binary bit is represented by 1 FlipFlop  n bits needs
n FlipFlops
 Combinational circuit: number of bits  circuit size
 To reduce memory size & circuit size  must reduce the number of bits 
What happens?
 Example: real number 10.69 = (?)2
1-bit
If m = 4, f = 4  10.69 adder
= (1010.1011) 2 = (10.6875)10  error = 0.025

If m = 4, f = 2  10.69 = (1010.10)2 = (10.5)10  error = 0.19

If m = 3, f = 4  10.69 = (010.1011)2 = (2.6875)10  Overflow!!

4-bit adder
38
Size of Number: Complexity vs. Performance

39
 Sent 10,000 bits  100 bit error

Sent 10,000 bits  1 bit error

40
 Contents
 Number System Introduction
Number systems used by human
Number systems used by computer

 Number System Conversion

 Signed Number Representation
 Fractional Number Representation
Fixed point number
Floating point number
 Size of Number: Complexity vs. Performance
41